Three-dimensional ordered-subset expectation maximization iterative protocol for evaluation of left ventricular...

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Cardiovascular Imaging

Background:

• Quantitative assessment of left ventricular (LV) function is crucial in cardiovascular diagnostics.
• Gated-SPECT (G-SPECT) using technetium-99m agents is a common imaging modality.
• Accurate quantification relies on effective image reconstruction algorithms.

Purpose of the Study:

• To evaluate the performance of a 3D ordered-subset expectation maximization (OSEM) algorithm for LV function quantification in G-SPECT.
• To optimize reconstruction parameters for the QGS program using a beating-heart phantom.
• To assess the impact of parameter variations on LV volume and ejection fraction measurements.

Main Methods:

• A dynamic heart phantom simulating LV function was imaged using G-SPECT.
• Tomographic reconstructions were performed using 3D OSEM with varying iterations (I), subsets (S), and Gaussian post-processing filter cutoff values (FWHM).
• Quantitative measurements of LV volumes and ejection fraction were obtained using the QGS program.

Main Results:

• The optimized 3D OSEM algorithm (10 iterations, 8 subsets, 10 mm FWHM) provided accurate LV ejection fraction (LVEF) calculations.
• LV volumes (EDV, ESV, SV) were consistently underestimated compared to known phantom volumes.
• Parameter variations within a defined range (80-160 EM-equivalent iterations, FWHM ≤10 mm) showed minimal clinically relevant differences in LVEF and volumes.

Conclusions:

• The QGS program, coupled with an optimized 3D OSEM algorithm, can accurately determine LVEF from G-SPECT data.
• LV volume underestimation by the algorithm requires consideration in clinical interpretation.
• The study identified optimal reconstruction parameters and demonstrated robustness to minor parameter adjustments for clinical applications.